Kober, Thomas and Weir, David (2015) Optimising agile social media analysis. In: 6th Workshop on Computational Approaches to Subjectivity, Sentiment and Social Media Analysis (WASSA 2015), 17th September 2015, Lisbon.
![]() |
PDF
- Published Version
Available under License Creative Commons Attribution No Derivatives. Download (493B) |
Abstract
Agile social media analysis involves building bespoke, one-off classification pipelines tailored to the analysis of specific datasets. In this study we investigate how the DUALIST architecture can be optimised for agile social media analysis. We evaluate several semi-supervised learning algorithms in conjunction with a Na ̈ıve Bayes model, and show how these modifications can improve the performance of bespoke classifiers for a variety of tasks on a large range of datasets.
Item Type: | Conference or Workshop Item (Paper) |
---|---|
Keywords: | social media analysis, sentiment analysis, twitter, opinion mining, text classification, semi-supervised learning, active learning, natural language processing, machine learning |
Schools and Departments: | School of Engineering and Informatics > Informatics |
Subjects: | Q Science > QA Mathematics > QA0075 Electronic computers. Computer science |
Related URLs: | |
Depositing User: | Thomas Kober |
Date Deposited: | 06 Jun 2016 13:57 |
Last Modified: | 06 Jun 2016 13:57 |
URI: | http://sro.sussex.ac.uk/id/eprint/61323 |
View download statistics for this item
📧 Request an update